[05:00] morning [05:56] heh, vendor.json keep flipping between 2 revisions of secboot package [06:58] jamesh: iirc pulseaudio tests fail pretty consistently on 20.10, i don't remember whether it's the same case for preseed though [06:58] mborzecki: thanks. [07:01] hmm interesting behavior of shellcheck https://github.com/koalaman/shellcheck/issues/650 [07:43] PR snapd#8851 closed: interface/fwupd: add more policies for making fwupd upstream strict [08:13] mborzeck1: hi, could you finish the review of #8959 when you have a moment, I will also look at it today [08:13] PR #8959: gadget,gadget/install: refactor partition table update [08:37] it'd be nice if for queued check jobs, there'd be some info since when the job has been in that state [08:38] PR snapd#9023 closed: sysconfig/cloudinit: add CloudInitStatus func + CloudInitState type [08:43] pedronis: do you think https://github.com/snapcore/snapd/pull/8861 is good to merge then? [08:43] PR #8861: data,packaging,wrappers: extend D-Bus service activation search path [08:45] jamesh: I need to look at it again, I'll try to do that today or tomorrow [08:45] thanks. === ricab_ is now known as ricab [10:08] PR snapd#8959 closed: gadget,gadget/install: refactor partition table update === Eighth_Doctor is now known as Conan_Kudo === Conan_Kudo is now known as Eighth_Doctor [13:41] zyga: not sure if you say, but I have a reproducer: https://bugs.launchpad.net/snapd/+bug/1888305 [13:41] Bug #1888305: layout lost after snapd refresh with long running content plugging snap [13:41] zyga: saw* [13:41] zyga: and hello :) [13:44] jdstrand: I think Zygmunt is off today [13:54] PR core20#76 opened: hooks/001-extra-packages.chroot: add gdbserver [13:56] PR core18#163 opened: hooks/001-extra-packages.chroot: add gdbserver [14:04] PR snapd#9039 opened: cmd/snap-seccomp/syscalls: add faccessat2 (2.45) [14:05] pedronis: ^^ [14:05] thanks [14:26] pedronis: fwiw google:ubuntu-20.10-64:tests/main/interfaces-audio-playback-record seems to be passing in isolation [14:27] interesting === alan_g_ is now known as alan_g [14:29] pedronis: the recording bit i handled by snap policy module which is added by ubuntu patches, and it queries the snapd api to get the connected plugs [14:30] yes [14:30] pedronis: there shouldn't be any races there though [14:34] pedronis: otoh, it ECONNREFUSED, so maybe pulseaudio just died there? [15:41] o/ [15:41] jdstrand: that's great news [15:41] hey zyga [15:42] * zyga woke up and catching up [15:42] today was not the best but it's almost over [15:42] ah [15:43] jdstrand: interesting that it is related to snapd refresh [15:43] it could be related to something i was looking at lately [15:45] jdstrand: https://forum.snapcraft.io/t/injecting-snapd-tools-into-base-snaps-and-keeping-them-up-to-date/12139 [15:45] (the last comment there) [15:45] no need to read it but I suspect that's really just uncovering a bug we knew existed for a while [15:45] it's great that there's a reproducer, I will try to pull myself together and write a spread test based on your super detailed bug report [15:46] * cachio lunch [15:47] hey ijohnson :) [16:08] ijohnson: I did a review of #9010 [16:08] PR #9010: cmd/snap-bootstrap/initramfs-mounts: call systemd-mount instead of the-tool [16:12] zyga: hi, I reviewed #8977 in my morning [16:12] PR #8977: cmd/snap: track started apps and hooks [16:13] thank you, I didn't look yet [16:13] * zyga looks [16:13] ah, thanks I'll send some updates there shortly [16:14] I will be around normally tomorrow [16:20] zyga: cool, thanks (though, I was told today you're off, so, only do it if it is fun and relaxing :) [16:20] jdstrand: yes because it's a mystery and I want to know :) [16:20] :) [16:21] zyga: hopefully for a spread test you will create a smaller reproducer snap ;) [16:23] yes, I just did [16:26] ijohnson: when you have time, #9030 needs your review [16:26] PR #9030: bootloader/assets: helpers for registering per-edition snippets, register snippets for grub [16:26] pedronis: yes I will review it today [16:27] and thanks for the review [16:27] np, thank you [16:33] mostly eod but I will look again later at the tests of some PRs to see if they are landable [16:44] jdstrand: btw, did you notice that the snap has this: target: $SNAP/data-dir/icons [16:44] twice [16:44] (I guess we should validate better) [16:47] zyga: ah, no, thanks! this doesn't use sounds so I never noticed it [16:49] what's the correct way to get the user's real home directory from within a snap? [16:51] Elleo: there isn't a perfect way but most people use /home/$LOGNAME [16:51] zyga: okay, thanks [16:51] I have a patch somewhere that provides that but it's not proposed anywhere [16:51] ah, okay [16:52] I vaguely remembered some discussion of adding an "$ORIG_HOME" variable or something, but wasn't sure if that ever happened [16:53] right, I don't think it did [17:48] ijohnson, hey [17:48] ijohnson, about the comment https://github.com/snapcore/snapd/pull/9027#discussion_r458208030 [17:48] PR #9027: tests: refresh/revert snapd in uc20 [17:49] ijohnson, cmatsuoka requested me to force the reboot when we refresh and revert snapd [17:49] to check if tpm works well afte the reboot [17:51] or it is ok and I should leave the code you proposed without explaining why we force the reboot [17:54] cachio: I understand [17:54] cachio: it's okay to still do a reboot after the refresh [17:54] cachio but the code as is is masking whether or not snapd triggered a reboot for the given type of refresh [17:54] ijohnson, yes, that could be a problem [17:55] The code you have always reboots and so if there was a bug where we trigger a reboot when we shouldn't be, or vice versa, your code would mask the problem and make the test pass [17:55] or perhaps it does not work properly if we dont reboot [17:55] Yes exactly [17:55] perhaps I could test both scnarios [17:56] That's why I wanted you to refactor it to check what kind of snap was being refreshed, and ensure the change for snaps that don't automatically trigger a reboot don't trigger a reboot [17:56] 1 forcing reboot and also without reboot [17:56] I provided some example code for how to do that [17:56] cachio: I don't think we need two tests [17:56] Just make the existing test slightly more accommodating to the different scenarios [17:57] ijohnson, I mean, same test test both thinks [17:57] things [17:58] ijohnson, but in the code you proposed it is going to reboot always for snapd and pc snaps [17:59] it is waiting until the refresh is comppleted fedore the reboot [17:59] right? [17:59] cachio: yes it will reboot anyways but if snapd triggered the reboot it would make one of the commands fail if I understand what I wrote correctly [17:59] Yes exactly if the reboot was triggered as part of the refresh then I think that command would exit non-xero and the test would fail [18:00] ijohnson, ah, ok, in that case I'll apply that [18:00] thanks for the suggestion [18:01] cachio: yaw, hopefully it's not too difficult to make work [18:01] trying now [18:21] PR snapd#9038 closed: tests: check for pids correctly [18:40] jdstrand: interesting, the reproducer fails with [18:40] + test-snapd-app.sh -c test -d /usr/share/hamster-applet [18:40] update.go:85: cannot change mount namespace according to change mount (/snap/gtk-common-themes/1506/share/gtk2/Materia-compact /snap/test-snapd-app/x1/data-dir/themes/Materia-compact none bind,ro 0 0): cannot write to "/snap/gtk-common-themes/1506/share/gtk2/Materia-compact" because it would affect the host in "/snap" [18:46] PR snapd#9040 opened: spread: add opensuse 15.2 and tumbleweed for qemu [18:46] zyga: huh [18:47] zyga: fyi, don't worry about tumbleweed lzo. I am doing it [18:49] k [18:49] that log is just a warning, not an error (in the sense that it doesn't stop) [18:49] but it's still wrong [18:49] yeah [18:50] * jdstrand idly wonders if his https://blog.strandboge.com/2019/04/16/cloud-images-qemu-cloud-init-and-snapd-spread-tests/ has been useful for others here... I've found it immensely useful for my spread testing :) [19:11] PR snapd#9018 closed: cmd/snap-preseed: check that target path exists and is a directory on --reset [19:19] kenvandine: hi [19:19] I found a small bug in gtk-common-themes [19:19] cat /snap/gtk-common-themes/current/meta/snap.yaml | grep Materia-dark-compact [19:19] this shows [19:19] - $SNAP/share/gtk2/Materia-dark-compact [19:20] but [19:20] ls -ld /snap/gtk-common-themes/1506/share/gtk2/Materia-dark-compact [19:20] ls: cannot access '/snap/gtk-common-themes/1506/share/gtk2/Materia-dark-compact': No such file or directory [19:20] jdstrand: the error was a red herring [19:31] PR snapd#9041 opened: osutil/group.go: treat all non-nil errs from user.Lookup{Group,} as Unknown* <⚠ Critical> [19:41] something is weird in a way that makes no sense [19:41] tell me about it [19:41] * ijohnson is slightly bitter about error checking getpwnam_r [20:08] zyga: have we just too many PRs, started too many tests, or something else is off, we have some PRs with queued tests since hours, for example: https://github.com/snapcore/snapd/pull/9012 [20:08] PR #9012: release/2.45: merge 2.45.2 release [20:21] pedronis: checking [20:22] only 4 spread runs in progress [20:22] now three [20:22] we may be queued on the github hosted tests [20:23] https://github.com/snapcore/snapd/actions?query=is%3Aqueued is very small [20:23] it's odd that there's a bunch of jobs that are still waiting though [20:24] for example https://github.com/snapcore/snapd/pull/9013 [20:24] PR #9013: many: merge 2.45.2 fixes back into master <⛔ Blocked> [20:25] actually now that I mention it that's the only one I can find other than the one pedronis found [20:25] hmmm [20:25] I did the close & reopen dance [20:25] now they show up in the queue [20:29] ijohnson: when it says "expected" [20:29] it means that it's not queued [20:29] but the rules say it should run [20:29] ah interesting [20:29] so likely a bug on their side [20:29] I close/reopend both [20:29] that reliably fixes it [20:29] zyga: so does it work to close and re-open someone else's PR? [20:29] or not? [20:29] yeah [20:29] I seem to have thought that wasn't allowed [20:30] it didn't change from queued [20:30] er [20:30] but maybe I just misremember that [20:30] from expected [20:31] I wonder if the fact it has conflicts matters [20:31] could it be that actions don't run [20:31] because it cannot be merged [20:31] so it cannot create the workspace for actions? [20:31] mmm interesting [20:31] could be [20:33] eh, shit [20:33] I just got a note that I need to re-do blood tests [20:33] for unknown reason [20:34] so I need to go there again and endure the morning without painkillers [20:34] eh, ... [20:34] more details in the morning [20:34] oh I'm so sorry to hear that [20:35] oh well [20:35] hope everything's still ok for your surgery [20:35] every day is something now [20:35] but it's just a few more days [20:35] then it's going to be better [20:36] week to get the stitches removed [20:36] about two weeks for most of the cut to heal [20:36] and in a month I should be able to sit normally again [20:54] jdstrand: do you have any experimental options set? [20:54] zyga: I do. is there an easy to list them all? [20:55] ah yes [20:55] $ sudo snap get system experimental [20:55] Key Value [20:55] experimental.hotplug true [20:55] experimental.parallel-instances true [20:55] zyga: ^ [20:56] thanks [21:37] jdstrand: no luck with a reproducer [21:37] I'll try more tomorrow [21:37] I pushed what I have [21:38] https://github.com/snapcore/snapd/compare/master...zyga:fix/lp-1888305?expand=1 [21:38] o/